Subzero: Don't "and" i1 values with 1.

Subzero: Don't "and" i1 values with 1.

In x86 lowering, i1 values are held in i8 register and memory slots.  We were conservatively "and"ing  them with 1 before zero-extending them for some lowering operations, but this "and" with 1 is unnecessary and just clutters the code.

We continue the invariant that all i1-produced values in an i8 slot are either 0 or 1.

BUG= https://code.google.com/p/nativeclient/issues/detail?id=4095
R=jpp@chromium.org

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a=commit;h=485d0773c9fd0605cc126e7e9020464a89ade564

[Jim Stichnoth, 2015-10-09 00:08:51]

stichnot@chromium.org changed reviewers:
+ jpp@chromium.org, kschimpf@google.com, sehr@google.com

[John, 2015-10-09 12:14:09]

lgtm

"link" this to the x86-32 target lowering bug.

LGTM otherwise.

https://chromiumcodereview.appspot.com/1394413002/diff/1/src/IceTargetLowerin...
File src/IceTargetLoweringX86BaseImpl.h (left):

https://chromiumcodereview.appspot.com/1394413002/diff/1/src/IceTargetLowerin...
src/IceTargetLoweringX86BaseImpl.h:2050: if (Src0RM->getType() == IceType_i1) {
are you 100% sure this is OK? I've wondered about this for a long time, but
given that llc does something similar, I stopped thinking about it.

[Jim Stichnoth, 2015-10-09 13:50:22]

https://chromiumcodereview.appspot.com/1394413002/diff/1/src/IceTargetLowerin...
File src/IceTargetLoweringX86BaseImpl.h (left):

https://chromiumcodereview.appspot.com/1394413002/diff/1/src/IceTargetLowerin...
src/IceTargetLoweringX86BaseImpl.h:2050: if (Src0RM->getType() == IceType_i1) {
On 2015/10/09 12:14:08, John wrote:
> are you 100% sure this is OK? I've wondered about this for a long time, but
> given that llc does something similar, I stopped thinking about it.

I'm about 99.44% sure...

The idea is to look through all operations that produce an i1 result and verify
that they all produce an 8-bit 0 or 1 (assuming any i1 inputs are also 0 or 1). 
For i1 arithmetic instructions, only add/or/xor are allowed, which is fine. 
Load is not allowed.  The trunc to i1 instruction still does "and 1".  So do
fptoui and fptosi.  Icmp and fcmp both do set<cc> instructions which result in 0
or 1.  So I think we're OK.

FWIW, the cross tests and spec2k are OK with this, which is what elevates my
confidence from 99% to 99.44%.

[Jim Stichnoth, 2015-10-09 13:52:24]

Committed patchset #1 (id:1) manually as
485d0773c9fd0605cc126e7e9020464a89ade564 (presubmit successful).